How do I get raw data off my wearable?

There are three routes. Some devices write files you can copy or export (Garmin FIT files, the Apple Health export.zip, ActiGraph or Axivity files). Most consumer brands offer a cloud API you reach with OAuth, through an official SDK or a community wrapper. Research devices and study platforms export CSV or raw sensor files. Consumer cloud APIs usually return summaries such as daily heart rate or sleep stages rather than the raw PPG waveform; raw waveforms generally need a device SDK that streams them, such as the Polar BLE SDK, or a research device.

What is the difference between a device SDK and a unified health-data API?

A device SDK or extractor talks to one brand. A unified API such as Open Wearables or Terra connects many brands and returns one schema, so you integrate once. Data access & tools lists 17 device extractors and SDKs and 9 unified APIs or schemas.

Which devices have the most tools?

Counting tools that name a brand: Garmin (14), Fitbit (8), Apple Watch / Apple Health (7), Oura (6) and Polar (6). Tools that work on exported signals without naming a device, such as NeuroKit2 or HeartPy, are not counted here.

Are these tools open source?

33 tools of 43 are open source. The rest are commercial services or source-available under an academic licence; each listing shows its licence as stated by the project.

How is this list kept up to date?

Every entry was checked against the tool's own repository, package page or website. Last-commit dates and archived status are refreshed weekly from each repository (last refresh 2026-11-09).

Wearable Signal Tools · Data access & tools

Data Access & Tools: Getting Raw Data Out of Wearables

43 verified ways to get data out of consumer and research wearables: 17 device data extractors and SDKs, 9 unified health-data APIs, 5 research collection platforms and 12 analysis toolkits. Filter by type, devices supported, open source, signal, language, licence and maintenance.

Four kinds of data access tools

Device data extractor / SDK 17: Libraries, SDKs and apps that pull data from one brand's device, app or cloud account, or read the files a device writes. For example: actipy, empatica-e4-local, fitdecode, Gadgetbridge.
Unified health-data API 9: One API or schema that normalises data from many wearable brands, so an app integrates once instead of once per vendor. For example: Junction (formerly Vital), Open mHealth schemas, Open Wearables, ROOK.
Research collection platform 5: Study platforms that enrol participants, collect wearable and phone data, track adherence and export it for analysis. For example: AWARE Framework (Android client), Beiwe, Fitabase, Labfront.
Analysis toolkit 12: Open-source packages that turn exported wearable signals into features, sleep, activity or HRV results. For example: FLIRT, GGIR, HeartPy, impSQI.

Which tool gets data off my device?

Pick your device under “Devices supported”, then check the type: an extractor or SDK for one brand, a unified API for many, a study platform for research cohorts, or an analysis toolkit once you have the data.

Questions about Data access & tools

How do I get raw data off my wearable?
There are three routes. Some devices write files you can copy or export (Garmin FIT files, the Apple Health export.zip, ActiGraph or Axivity files). Most consumer brands offer a cloud API you reach with OAuth, through an official SDK or a community wrapper. Research devices and study platforms export CSV or raw sensor files. Consumer cloud APIs usually return summaries such as daily heart rate or sleep stages rather than the raw PPG waveform; raw waveforms generally need a device SDK that streams them, such as the Polar BLE SDK, or a research device.
What is the difference between a device SDK and a unified health-data API?
A device SDK or extractor talks to one brand. A unified API such as Open Wearables or Terra connects many brands and returns one schema, so you integrate once. Data access & tools lists 17 device extractors and SDKs and 9 unified APIs or schemas.
Which devices have the most tools?
Counting tools that name a brand: Garmin (14), Fitbit (8), Apple Watch / Apple Health (7), Oura (6) and Polar (6). Tools that work on exported signals without naming a device, such as NeuroKit2 or HeartPy, are not counted here.
Are these tools open source?
33 tools of 43 are open source. The rest are commercial services or source-available under an academic licence; each listing shows its licence as stated by the project.
How is this list kept up to date?
Every entry was checked against the tool's own repository, package page or website. Last-commit dates and archived status are refreshed weekly from each repository (last refresh 2026-11-09).

Listings, summaries, licences and commercial-use fields are shown as recorded for each project and checked against each tool’s own site or repository, as of Oct 21, 2026. Maintenance is computed from the date of the last commit; “Archived upstream” means the project’s repository is marked archived. Check the licence in each project’s repository before you use it.